Working Paper: CEPR ID: DP6165
Authors: Arianna Degan; Antonio Merlo
Abstract: In this paper we address the following question: To what extent is the hypothesis that voters vote sincerely testable or falsifiable? We show that using data only on how individuals vote in a single election, the hypothesis that voters vote sincerely is irrefutable, regardless of the number of candidates competing in the election. On the other hand, using data on how the same individuals vote in multiple elections, the hypothesis that voters vote sincerely is potentially falsifiable, and we provide general conditions under which the hypothesis can be tested. We then consider an application of our theoretical framework and assess whether the behaviour of voters is consistent with sincere voting in U.S. national elections in the post-war period. We find that by and large sincere voting can explain virtually all of the individual-level observations on voting behaviour in presidential and congressional U.S. elections in the data.
Keywords: falsifiability; spatial models; testing; voting
JEL Codes: C12; C63; D72
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Single election data (K16) | observed vote is rationalized as sincere voting (D79) |
| Multiple elections (D72) | meaningful test for sincere voting (K16) |
| Sincere voting hypothesis (D79) | explains individual-level voting behavior (D72) |
| Failure to account for voting behavior (D72) | measurement errors (C20) |